Papers by Xuhui Zhou
COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements (2023.findings-acl)
Copied to clipboard
Xuhui Zhou, Hao Zhu, Akhila Yerukola, Thomas Davidson, Jena D. Hwang, Swabha Swayamdipta, Maarten Sap
| Challenge: | Understanding the harms and offensiveness of statements requires reasoning about the social and situational context in which they are made. |
| Approach: | They propose a context-aware formalism for explaining the intents, reactions, and harms of offensive statements grounded in their social and situational contexts. |
| Outcome: | The proposed framework is the first context-aware formalism for explaining the intents, reactions, and harms of offensive statements grounded in their social and situational context. |
Enhancing Extractive Question Answering in Multiparty Dialogues with Logical Inference Memory Network (2025.coling-main)
Copied to clipboard
| Challenge: | Existing models for multiparty dialogue question answering (QA) do not consider logical inference relations in multiparty dialogs, leading to suboptimal performance. |
| Approach: | They propose a memory network with logical inference for extractive QA in multiparty dialogues. |
| Outcome: | The proposed model achieves state-of-the-art on Molweni and FriendsQA benchmarks. |
DPDLLM: A Black-box Framework for Detecting Pre-training Data from Large Language Models (2024.findings-acl)
Copied to clipboard
Baohang Zhou, Zezhong Wang, Lingzhi Wang, Hongru Wang, Ying Zhang, Kehui Song, Xuhui Sui, Kam-Fai Wong
| Challenge: | Existing methods to detect pretraining data from large language models are unrealistic to them. |
| Approach: | They propose to detect pre-training data from LLM in a black-box way by using GPT-2 as reference model and feed it with sequence probabilities to detect whether it was used to train it. |
| Outcome: | The proposed framework outperforms existing methods on the benchmark datasets and shows that it is effective on different popular LLMs. |
Is this the real life? Is this just fantasy? The Misleading Success of Simulating Social Interactions With LLMs (2024.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in large language models have enabled richer social simulations . however, the role of information asymmetry in these simulations has been overlooked . |
| Approach: | They develop an evaluation framework to simulate social interactions with LLMs in different settings. |
| Outcome: | The proposed framework performs better in unrealistic, omniscient simulation settings but struggles in those with information asymmetry. |
FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing evaluations for theory of mind (ToM) use passive narratives that lack interactivity. |
| Approach: | They propose a benchmark to stress-test ToM within information-asymmetric conversational contexts via question answering. |
| Outcome: | The proposed benchmark is challenging for state-of-the-art language models, which perform significantly worse than humans even with chain-of thought reasoning or fine-tuning. |
MCIL: Multimodal Counterfactual Instance Learning for Low-resource Entity-based Multimodal Information Extraction (2024.lrec-main)
Copied to clipboard
| Challenge: | Existing methods to perform multimodal information extraction only investigated entity-based tasks under supervised learning with adequate labeled data. |
| Approach: | They propose to investigate the entity-based MIE tasks under the low-resource settings by decomposing the features into image, entity, and context factors. |
| Outcome: | The proposed method is able to perform on two public MIE benchmark datasets and the experimental results confirm it. |
PM2F2N: Patient Multi-view Multi-modal Feature Fusion Networks for Clinical Outcome Prediction (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods focused on time series data but ignored clinical notes . fusion of multi-modal features of patients from different views is not feasible due to the time series and clinical notes data being stored as time series. |
| Approach: | They propose to combine time series and clinical notes to fuse multi-modal features of patients from different perspectives using graph neural networks. |
| Outcome: | The proposed method is superior to existing models on MIMIC-III benchmark. |
Challenges in Automated Debiasing for Toxic Language Detection (2021.eacl-main)
Copied to clipboard
| Challenge: | Existing methods for debiasing toxic language data are limited in their ability to prevent biased behavior in toxic language detection systems. |
| Approach: | They propose to debiase toxic language detection models using lexical and dialectal markers using synthetic labels instead of traditional methods. |
| Outcome: | The proposed method reduces dialectal associations with toxicity despite the use of synthetic labels . |
BIG5-CHAT: Shaping LLM Personalities Through Training on Human-Grounded Data (2025.acl-long)
Copied to clipboard
| Challenge: | Existing methods for embedding human personality traits into LLMs are limited by realism and validity issues. |
| Approach: | They propose to use a large-scale dataset to embed human personality traits into LLMs . they use supervised fine-tuning and direct preference optimization to train LLM models . |
| Outcome: | The proposed methods outperform prompting on personality assessments and IPIP-NEO, and show higher conscientiousness, agreeableness, lower extraversion, and lower neuroticism on reasoning tasks. |
Multilevel Text Alignment with Cross-Document Attention (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing alignment methods operate at a single, predefined level and cannot learn to align texts at sentence and document levels. |
| Approach: | They propose a learning approach that equips hierarchical attention encoders for representing documents with a cross-document attention component, enabling structural comparisons across different levels. |
| Outcome: | The proposed model outperforms existing hierarchical, attention encoders on citation recommendation and plagiarism detection tasks. |
Selecting Key Views for Zero-Shot Entity Linking (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Entity linking is a task of assigning ambiguous mentions in textual input to entities in knowledge bases. |
| Approach: | They propose a framework to align mentions in text to entities in knowledge bases . they use unsupervised clustering to select key views from descriptions . |
| Outcome: | The proposed framework achieves state-of-the-art on the zero-shot entity linking dataset. |
Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Models (2024.eacl-long)
Copied to clipboard
Natalie Shapira, Mosh Levy, Seyed Hossein Alavi, Xuhui Zhou, Yejin Choi, Yoav Goldberg, Maarten Sap, Vered Shwartz
| Challenge: | Recent work suggests that Large Language Models (LLMs) exhibit Neural Theory-of-Mind (N-ToM) however, prior work reached conflicting conclusions regarding those abilities. |
| Approach: | They examine the extent of Large Language Models’ N-ToM abilities through an extensive evaluation of 6 tasks and find that LLMs struggle with adversarial examples . |
| Outcome: | The proposed metrics show that LLMs exhibit certain N-ToM abilities, but this behavior is far from robust. |
Imperfectly Cooperative Human-AI Interactions: Comparing the Impacts of Human and AI Attributes in Simulated and User Studies (2026.findings-acl)
Copied to clipboard
Myke C. Cohen, Mingqian Zheng, Neel Bhandari, Hsien-Te Kao, Xuhui Zhou, Daniel Nguyen, Laura Cassani, Maarten Sap, Svitlana Volkova
| Challenge: | In simulations, personality traits and AI attributes were comparatively influential, but with actual human subjects, AI attributes – particularly transparency – were much more impactful. |
| Approach: | They compare a purely simulated dataset and a parallel human subjects experiment to examine how human personality traits and AI design characteristics jointly shape interaction outcomes in imperfectly cooperative scenarios. |
| Outcome: | The results show that personality traits and AI attributes are comparatively influential in simulations, but with actual human subjects, they are much more impactful. |
Words Like Knives: Backstory-Personalized Modeling and Detection of Violent Communication (2025.emnlp-main)
Copied to clipboard
| Challenge: | a recent study examines the role of personalization in enabling LLMs to serve as effective mediators in human communication for authentic connection. |
| Approach: | They leverage nonviolent communication theory to evaluate LLMs in detecting conversational breakdowns . they annotate a subset of dialogues and obtain fine-grained labels of communication breakdown types . |
| Outcome: | The proposed dataset analyzes human interactions and relationships in a human context. |
TimeR4 : Time-aware Retrieval-Augmented Large Language Models for Temporal Knowledge Graph Question Answering (2024.emnlp-main)
Copied to clipboard
| Challenge: | Temporal Knowledge Graph Question Answering (TKGQA) aims to answer temporal questions using knowledge in Temporal knowledge graphs (TKTs). |
| Approach: | They propose a Time-aware retrieve-rewrite-retrieve-rerank framework to integrate temporal knowledge from TKGs into Large Language Models (LLMs) to reduce temporal hallucination, they propose rewrite module to rew questions using background knowledge stored in TKG's, then implement a retrieve-rank module to retrieve semantically and temporally relevant facts from Tkgs and rerank them according to temporal constraints. |
| Outcome: | The proposed approach achieves relative gains of 47.8% and 22.5% on two datasets, underscoring its effectiveness in boosting the temporal reasoning abilities of LLMs. |
AI-LieDar : Examine the Trade-off Between Utility and Truthfulness in LLM Agents (2025.naacl-long)
Copied to clipboard
| Challenge: | LieDar is a framework to study how LLM-based agents navigate these scenarios in a multi-turn interactive setting. |
| Approach: | They propose a framework to study how LLM-based agents navigate these scenarios in an interactive multi-turn setting. |
| Outcome: | The proposed framework shows that all models are truthful less than 50% of the time, although truthfulness and goal achievement rates vary across models. |
Beyond Timestamps: Bridging Forward and Backward Reasoning in Temporal Numerical and Relational Understanding (2026.acl-long)
Copied to clipboard
| Challenge: | Existing benchmarks for Temporal Numerical and Relational reasoning rely on single-task evaluation paradigms. |
| Approach: | They propose a benchmark to evaluate Temporal Numerical and Relational reasoning . they propose QA and verification, and a Consistency Rate to quantify robustness . |
| Outcome: | The proposed framework evaluates both Temporal Numerical and Relational reasoning . it measures the alignment between QA and FV and the Consistency Rate measures robustness across these directions. |
MELOV: Multimodal Entity Linking with Optimized Visual Features in Latent Space (2024.findings-acl)
Copied to clipboard
| Challenge: | Existing approaches to multimodal entity linking focus on textual contexts but lack in social media vision modality. |
| Approach: | They propose a latent space vision feature optimization framework MELOV to address these challenges . they exploit variational autoencoder to mine shared information and generate text-based visual features . |
| Outcome: | The proposed framework is superior to existing methods on three benchmark datasets. |
RPD: A Distance Function Between Word Embeddings (2020.acl-srw)
Copied to clipboard
| Challenge: | Existing word embeddings are poorly understood, but little is known about how they differ between different sets of word embeds. |
| Approach: | They propose a metric called Relative Pairwise Inner Product Distance to quantify the distance between different word embeddings. |
| Outcome: | The proposed metric measures the distance between different sets of embeddings and investigates the influence of different training processes and corpora. |
Don’t Take This Out of Context!: On the Need for Contextual Models and Evaluations for Stylistic Rewriting (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing stylistic text rewriting methods ignore the context of the text, causing generic, incoherent, and generic outputs. |
| Approach: | They propose a contextual evaluation metric that integrates preceding context into stylistic text rewriting. |
| Outcome: | The proposed metric integrates the preceding textual context into rewriting and evaluation stages . human preferences are better reflected by the proposed criterio and other metrics . |
BioFEG: Generate Latent Features for Biomedical Entity Linking (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing approaches to biomedical entity linking suffer from multiple types of errors due to the rarity of many biomedically relevant entities in real-world scenarios. |
| Approach: | They propose a latent feature generation framework to generate latent semantic features for unseen entities to capture fine-grained coherence information of unseened entities. |
| Outcome: | The proposed framework is superior to existing models on two benchmark datasets. |
Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection (2022.naacl-main)
Copied to clipboard
| Challenge: | toxicity annotations are often ignored because of its subjective nature and lack of nuance. |
| Approach: | They examine the effect of annotator identities and beliefs on toxic language annotations by considering posts with three characteristics: anti-Black language, African American English (AAE) dialect, and vulgarity. |
| Outcome: | The findings show strong associations between annotator identity and beliefs and ratings of toxicity. |
Improving Zero-Shot Entity Linking Candidate Generation with Ultra-Fine Entity Type Information (2022.coling-1)
Copied to clipboard
| Challenge: | Entity linking is a task of assigning entity mentions to referent entities in a knowledge base. |
| Approach: | They propose to use ultra-fine-grained type information to improve the generalization ability of EL models by utilizing a low-level task to extract ultra-finish entity type information. |
| Outcome: | The proposed model achieves state-of-the-art in the zero-shot entity linking task . |
DataArc-SynData-Toolkit: A Unified Closed-Loop Framework for Multi-Path, Multimodal, and Multilingual Data Synthesis (2026.acl-demo)
Copied to clipboard
Zhichao Shi, Cehao Yang, Hao Zhou, Xiaojun Wu, Huajie Li, Xuhui Jiang, Chengjin Xu, Yuanzhuo Wang, Jian Guo
| Challenge: | Existing synthetic data tools are limited by convoluted workflows, fragmented data standards, and limited scalability across modalities. |
| Approach: | They develop an open-source framework that aims to reduce the technical barrier to synthetic data generation and subsequent model training. |
| Outcome: | The proposed framework achieves an optimal balance between generation efficiency and data quality. |